Search results

[Bugs] Second batch of verified findings from the unit-coverage initiative (fold_weight crash on Llama-4/GPT-OSS MoE, KD-loss default footgun, and smaller items)

Status: Open.

#1926 In NVIDIA/Model-Optimizer;

· arham766 opened on Jul 6, 2026

[Coverage] Direct unit-test suites for five untested core modules (conversion, core_utils, tensor_quantizer, model_calib, export/postprocess) — plus bugs found along the way

Status: Open.

#1902 In NVIDIA/Model-Optimizer;

· arham766 opened on Jul 5, 2026

Run downstream eval, e.g., mmlu, every n iterations during mbridge distillation

featureIssue in actual ModelOpt featureIssue in actual ModelOpt feature

feature requestNew feature or requestNew feature or request

torch.distillation

waiting for feedback

Status: Open.

#1896 In NVIDIA/Model-Optimizer;

· danielkorzekwa opened on Jul 3, 2026

warnings.warn(f"RealQuantLinear: No real-quant GEMM found: {self}.")

bugSomething isn't workingSomething isn't working

model support

torch.quantization

waiting for feedback

Status: Open.

#1895 In NVIDIA/Model-Optimizer;

· zewenli98 opened on Jul 2, 2026

Create create_data_blend utility

feature requestNew feature or requestNew feature or request

Status: Open.

#1876 In NVIDIA/Model-Optimizer;

· danielkorzekwa opened on Jul 1, 2026

NVFP4 quantized (shared) lm_head fails fake-quant in training forward: ScaledE4M3Function 'Only support E=4 & M=3'

bugSomething isn't workingSomething isn't working

torch.quantization

waiting for feedback

Status: Open.

#1875 In NVIDIA/Model-Optimizer;

· yeyu-nvidia opened on Jul 1, 2026

Clarify canonical documentation between guides and example READMEs (pruning)

bugSomething isn't workingSomething isn't working

documentationImprovements or additions to documentationImprovements or additions to documentation

feature requestNew feature or requestNew feature or request

waiting for feedback

Status: Open.

#1871 In NVIDIA/Model-Optimizer;

· danielkorzekwa opened on Jul 1, 2026

Support for INT4_AWQ quantization for Nemotron-3-Nano-30B-A3B

bugSomething isn't workingSomething isn't working

fw-megatron

torch.quantization

Status: Open.

#1865 In NVIDIA/Model-Optimizer;

· ADharaUTEXAS123007 opened on Jun 30, 2026

fsdp2_aware_weight_update hides the real error with an UnboundLocalError

bugSomething isn't workingSomething isn't working

torch.quantization

waiting for feedback

Status: Open.

#1859 In NVIDIA/Model-Optimizer;

· dinhxuanvu opened on Jun 30, 2026

Do not getting access to Nemotron-Pretraining-SFT-v1 (needed for a mintron distillation tutorial)

questionHelp is is neededHelp is is needed

waiting for feedback

Status: Open.

#1852 In NVIDIA/Model-Optimizer;

· danielkorzekwa opened on Jun 29, 2026

Utility to tokenize a dataset for distillation after pruning (puzzletron/minitron) based on DATA_BLEND variable

feature requestNew feature or requestNew feature or request

torch.pruning

Status: Open.

#1851 In NVIDIA/Model-Optimizer;

· danielkorzekwa opened on Jun 29, 2026

Provide better way of recommending a dataset for global distillation for puzzletron algorithm

feature requestNew feature or requestNew feature or request

torch.pruning

Status: Open.

#1850 In NVIDIA/Model-Optimizer;

· danielkorzekwa opened on Jun 29, 2026